He led safeguarding at a major AI company. His next career move will surprise you.

/

Mrinank Sharma, who had led Anthropic’s safeguards research team since its launch last year, shared his resignation letter in a post on X Monday morning, which quickly garnered attention and has been viewed 1 million times. In his letter, Sharma said the “world is in peril,” not just from AI, but a “whole series of interconnected crises unfolding in this very moment.” After leaving Anthropic, Sharma said he may pursue a poetry degree and “devote myself to the practice of courageous speech,” adding he wants to “contribute in a way that feels fully in my integrity.”

Read more at Forbes

16 Articles about Social Media & AI

/

AI Definitions: Convolutional neural networks 

/

Convolutional neural networks (CNNs or ConvNet) – These deep learning artificial neural networks, often used in computer vision for object recognition, are trained on thousands of images. It works similarly to how our human eye processes images. The network is trained to recognize "kernels," which are tiny pieces of an image. However, they can fail when they encounter the same objects under new lighting conditions or from a different angle. CNNs play a role in unlocking our phones with our faces, identifying road signs in self-driving cars, and automatically tagging people in our photo galleries. CNNs were first introduced in 1989 by NYU professor Yann LeCun and have been used with autonomous vehicles and security camera systems.

More AI definitions

16 Articles about Students Using & Impacted by AI

/

AI Definitions: Agentic AI

/

Agentic AI – Able to operate more independently than AI Agents, Agentic AI operates like a workflow, able to adjust its strategy and continuously learn as it encounters different situations. Agentic AI systems aren't passive tools waiting for input or mere automation. They can update plans based on intermediate findings without needing continuous human supervision. It’s not just following the rules as agents do, Agentic AI is supposed to be a colleague that can analyze a problem, propose a plan, and take action. Think of agentic AI as a team of digital colleagues where some agents are coordinators and some are specialists. It might call out to additional models or external systems, such as a search engine or querying a database to complete a task. This can be particularly effective in data-heavy fields such as biology, chemistry, and drug discovery. On a personal level, instead of simply helping you find a hotel room to book, agentic AI can plan the trip if it is given access to programs with your schedule and preferences. Agents can better handle the back-and-forth interactions that most real workflows require than rule-based systems. Despite its capabilities, AI agents can struggle in open-ended or unpredictable environments, especially when tasks lack clear structure or context. It will likely take years to for most agentic AI systems to be tailored to specific industries or problems.

More AI definitions

How Many Friends Can We Have?

/

The number of people with whom we can maintain a stable relationship is about 150, according to British anthropologist Robin Dunbar.

He says: “We devote around 40 percent of our available social time to our 5 most intimate friends and relations … and the remaining 60 percent in progressively decreasing amounts goes to the other 145.”

Friendship is the single most important factor influencing our health, well-being, and happiness. Creating and maintaining friendships is, however, extremely costly, in terms of both the time that has to be invested and the cognitive mechanisms that underpin them. Part of friendship is the act of mentalizing, or mentally envisioning the landscape of another's mind. Cognitively, this process is extraordinarily taxing, and as such, intimate conversations seem to be capped at about four people before they break down and form smaller conversational groups.  

Read more at the BigThink

9 Things People are Trying to Get AI to Do

/

Rent a Human

/

Rent a Human.ai is a new site promoted as a place to book humans for real-world tasks your AI can’t do. “According to the site, more than 81,000 "rentable humans" have already signed up to offer paid services to bots. The tasks themselves range from mundane errands like picking up packages to holding signs or delivering flowers to Anthropic. Rent-a-Human requires users to connect crypto wallets in order to get paid.” More at Mashable

AI Definitions: GPT

/

GPT (Generative Pre-trained Transformer) – GPT refers to a LLM (large language model) type of AI that first goes through an unsupervised period (no data labeling by humans) followed by a supervised "fine-tuning" phase (some labeling). G is for Generative because it generates words. P is for Pre-trained because it’s trained on a lot of text. This step is called pre-training because many language models (like the one behind ChatGPT) go through important additional stages of training known as fine-tuning to make them less toxic and easier to interact with. T stands for Transformer which is a relatively recent breakthrough in how neural networks are wired. They were introduced in a 2017 paper by Google researchers, and are used in many of the latest AI advancements, from text generation to image creation.

More AI definitions

Is your camera documenting reality – or negotiating with it?

/

A Reddit user held a phone up to a deliberately blurry, pixelated image of the Moon on his computer. Happy to oblige, his phone snapped a nice clear picture, full of craters and shadows which didn't actually appear in the original photo. The reality is that AI will recognise the Moon and fill in details when the camera can't pick them up. It's called computational photography. Your phone goes far beyond collecting the light that hits your camera's sensors. It's guessing what the image would look like if the camera was better and then building it for you, he says. The next time you take a photo, ask yourself, is your camera documenting reality – or negotiating with it? -BBC

17 Articles about Using AI

/

The most durable advantage in a world of abundant machine intelligence

/

In a world of abundant machine intelligence, the most durable advantage will be broad intellectual range. As routine analysis becomes automated, what distinguishes professionals is the ability to synthesize across domains, to see patterns that specialists miss, to exercise judgment. The best candidates think independently, navigate ambiguity without waiting for instruction, analyze the questions that were not asked but should have been and own their decisions. They use A.I. — as a tool but not a crutch. Where evidence is mixed and incomplete, professionals must possess the skills to make things better where machines cannot. - Blair Effron writing in The New York Times